[184] Can ER and CK5 Identify Atypical Epithelial Proliferations in Papillary Lesions of the Breast on Core Biopsy?

AN Grin, FP O'Malley, AM Mulligan. University of Toronto, Toronto, ON, Canada; Mount Sinai Hospital, Toronto, ON, Canada; St. Michael's Hospital, Toronto, ON, Canada

Background: The presence of atypical or usual epithelial proliferations within papillary breast lesions complicates their interpretation on core biopsy. Although histologic features are paramount, we evaluated the combination of estrogen receptor (ER) and cytokeratin 5 (CK5) as an aid in the distinction of usual duct hyperplasia (UDH) from atypical proliferations in this setting.
Design: Core biopsies from 185 papillary lesions were reviewed and of these,82 cases were selected for immunohistochemical study based on the presence of an epithelial proliferation between the fibrovascular cores. Fifty-two cases were used as the test set and 30 cases, with subsequent surgical excision, were used as the validation set. The epithelial proliferation was evaluated for staining intensity and percentage of positive cells using CK5 and ER.
Results: Of the test set, a consensus diagnosis of benign papilloma with UDH was reached in 20 cases and atypical papillary lesion (either limited (ADH) or more extensive (DCIS)) in 32 cases. Expression of both CK5 and ER was significantly different in benign lesions when compared with atypical lesions (p<0.0001). Benign lesions typically showed an ER-low/CK5-high profile and atypical lesions showed an ER-high/CK5-low profile with ER-high expression defined as diffuse strong staining in >90 % of cells. Anything less or variation in intensity was considered ER-low. CK5-high expression was defined as a mosaic pattern of staining in >20% of cells and CK5-low as absent or staining in < 20% of cells. Based on their staining profile 29 of the 30 validation cases were correctly classified using the excision specimen as the gold standard (Table 1). The remaining case did not fit either the benign or atypical staining profile.

Table 1.Staining profile of validation set according to diagnosis on excision specimen.

Excision diagnosis	ER-high/CK5-low N (%)	ER-low/CK5-high N (%)	ER-low/CK5-low N (%)	ER-high/CK5-high N (%)
Benign	0	15/15 (100%)	0	0
Atypical/DCIS/Invasive	14/15 (93%)	0	1/15 (7%)	0

Atypical profile = 93% sensitive, 100% specific;Benign profile = 100% sensitive, 100% specific.


Conclusions: Patterns and extent of ER and CK5 staining, when used together, are valuable adjunct stains to differentiate UDH from atypical proliferations within papillary lesions on core biopsy. This is the first report to illustrate the value of the combined use of ER and CK5 in the characterization of epithelial proliferations of breast.
Category: Breast

Monday, March 9, 2009 9:30 AM

Poster Session I Stowell-Orbison/Autopsy Award # 32, Monday Morning